VAPOR PERMEABILITY PROVIDES NO …

1 1 VAPOR PERMEABILITY PROVIDES NO performance BENEFIT FOR ROOFING UNDERLAYMENTS IN VENTILATED ATTICST esting shows breathable underlayment PROVIDES little or no moisture transfer value in asphalt roofi ng Joseph Lstiburek, , P. Eng., ASHRAE Fellow, Building Science CorporationDr. Achilles Karagiozis, Director of Building Science, Owens CorningPaul Gassman, P. Eng., Product Engineer, Owens Corning October 20112 AbstractMany manufacturers have introduced synthetic underlayments in the roofi ng market to serve as secondary water shedding barriers under roof shingles. Traditional organic asphalt felt has served this purpose for years, but durability has diminished over time and the product is inferior to recently available synthetic products made with layers of composite polyolefi n. While traditional felts are permeable, allowing moisture VAPOR transfer over time, newer synthetic materials are typically nonpermeable, allowing very little moisture transport.

2 Recent testing and evaluation at Owens Corning demonstrates that adding breathability to synthetic underlayment PROVIDES no advantage to the building performance of an asphalt roof assembly. A modifi ed ASTM E 96 Dry Cup testing method demonstrated that standard overlapping shingle construction creates its own VAPOR barrier system, preventing both the transport of moisture from exterior weather elements, as well as preventing moisture VAPOR escape from the building interior. Moisture transfer through the roof cannot be achieved simply by making the underlayment material breathable. Because the experimental data indicates that the multi-layered shingle system creates a VAPOR barrier, a properly designed and installed attic ventilation system, or a properly designed and installed unvented roof assembly1, is necessary to protect the roof sheathing from moisture within the focus of this paper is to investigate the system performance of standard asphalt shingles and to evaluate the impact of installing nonbreathable underlayments between the shingle layer and the roof deck.

3 The research performed indicates that nonbreathable underlayments may be installed below asphalt shingle roofi ng materials with comparable or better moisture focus of this paper is to investigate the system performance of standard asphalt shingles and to evaluate the impact of installing nonbreathable underlayments. Figure 1: Typical Underlayment Installation1 For Owens Corning Roofi ng and Asphalt, LLC s position on Unvented Attics see Technical Services Bulletin RD-01012011. Request by phone: 1-419-248-6557 or e-mail: roofi ng market has seen an onslaught of new synthetic underlayment products in the past 10 years. These products bring many advantages to the installer: increased speed of installation; lighter weight; and signifi cantly stronger physical characteristics, resulting in increased wind uplift performance than typical asphalt felt underlayment. These products also benefi t homeowners by protecting their homes as a durable moisture barrier from the elements over an extended period of dry-in during construction or re-roofi synthetic products claim to improve performance of the roofi ng system by including breathability as an added feature.

4 But does this feature truly add any benefi t to a typical asphalt shingle roofi ng system? In a typical installation, underlayment is sandwiched between the plywood or oriented strand board (OSB) roof deck and a covering layer of asphalt shingles. Does a breathable underlayment allow attic moisture to escape? Does it allow a roof deck to breathe? These questions are addressed in this white synthetic products claim to improve performance of the roofi ng system by including breathability. 42 International Residential Code for One- and Two-family Dwellings, First Printing: March 2009, Copyrighted 2009 by International Code Council, Inc., Publications, 4051 West Flossmoor Road, Country Club Hills, IL ASTM E 2: #15 Felt Underlayment Sample ASTM E 96 6" dia. Dry Cup Defi ning BreathabilityThe 2009 International Residential Code (IRC) defi nes a VAPOR permeable membrane as a material or covering having a permeance rating of 5 perms or greater, when tested in accordance with the desiccant method with Procedure A of ASTM E 96.

5 A VAPOR permeable material permits the passage of moisture VAPOR . 2 However, traditional convention within the building industry defi nes: a material with a perm of less than as VAPOR impermeable a material with a perm of between and as VAPOR semi impermeable a material with a perm of between and as VAPOR semi permeable and a material with a perm of greater than 10 as VAPOR permeableAdditionally, a VAPOR barrier is defi ned as less than perm and a VAPOR retarder is defi ned as less than perm. As such, some inconsistency exists between the IRC and traditional PermeabilityThe industry standard test method for water VAPOR transmission, also known as permeance, is ASTM E 96 Standard Test Methods for Water VAPOR Transmission of Because the IRC recommends the Dry Cup process defi ned as Procedure A, and with consideration of the sample size required to examine the overall system effectively, the dry or desiccant method was used for the shingle roofi ng system testing in this test is relatively straightforward.

6 Test material is sealed over a container of desiccant and placed in a humidity- and temperature-controlled chamber. Over time, the desiccant will draw moisture from the ambient air in the chamber through the test material, which is then trapped in the desiccant. Measuring the water weight-gain in the sealed container over time gives a value for permeance, which measures the time rate of water VAPOR transmission through the test industry standard test method for water VAPOR transmission, also known as permeance, is ASTM E 96 Standard Test Methods for Water VAPOR Transmission of Materials. 5 Individual Component TestingASTM E 96 enables the measurement of individual materials as well as the assembled system. Individual material testing was accomplished using 6" cups with a wax seal ring around the perimeter to close the sample cup. Table 1 PROVIDES a baseline understanding of the roofi ng system Installation InstructionsOwens Corning Classic 3-tab shingles were used in all testing.

7 Application instructions for these shingles include a 5" vertical exposure on the 12" high shingle and a 6" offset on the horizontal dimension for shingle lapping. This is the industry standard practice for shingle method PROVIDES an overlapping water shedding construction necessary to keep rainwater out. The same principle greatly increases the travel path or fl ow length for air movement through the same assembly. This resistance to airfl ow is likely the greatest contributing factor in creating the VAPOR resistance which this testing a 12" height and a 5" exposure on the individual shingles, the overlapping system results in an air path which always has a double layer of shingles, and a triple layer of material at each vertical intersection for air and moisture VAPOR to migrate through. The 36" width of the shingles also introduces a complicated path for any air and moisture VAPOR to travel through in the horizontal or lateral direction.

8 Additionally, shingles are relatively heavy and fl at and have a rough surface, all three of which are physical characteristics that increase the resistance to airfl ow in the roofi ng to airfl ow is likely the greatest contributing factor in creating the VAPOR resistance which this testing demonstrates. TEST MATERIALPERMEANCE RATINGA sphalt Shingles #15 16" OSB Decking 1: Typical Single Component Testing44 ASTM E 96 Sect. 11, Procedure for Desiccant 3: Overlapping Shingle Construction6 Roofi ng Shingle Layer TestingA 24" wide by 36" long commercially available plastic pan was selected as the test dish for the ASTM E 96 testing on large-scale system components. This size allowed for a typical asphalt shingle application with seven overlapping horizontal rows of shingles, and three vertical butt joints between adjacent shingles. To simulate the shingle layer during application, an OSB frame was fabricated with large slotted windows to allow moisture VAPOR transfer through this layer with minimal interference from the OSB, while the OSB allowed the shingle attachment as per manufacturer s testing on this application showed the multi-layered asphalt shingle system when installed per the manufacturer s instructions has an average moisture VAPOR transfer rate of perms.

9 Demonstrating a measured perm of less than showed that moisture transfer through the multi-layer asphalt shingles is negligible. This important fact begs the question: If the asphalt shingles act as a VAPOR retarder on the roof, then what value is added with the introduction of a breathable roof underlayment below it? Any moisture within the roof deck or the attic will not be able to move through the roofi ng system, regardless of the PERMEABILITY of the roofi ng underlayment layer. A measured perm of less than showed that moisture transfer through the multi-layer asphalt shingles is negligible. 7 Computer AnalysisUsing a state-of-the-art hygrothermal (combined heat and moisture transport) model developed by Karagiozis et al (2001), two simulations were performed representing the impermeable and permeable underlayment. This model has been validated for a number of wall and roof systems showing good agreement with fi eld data by a number of organizations.

10 These two simulations were performed to investigate the moisture storage differences between the two types of roof underlayment approaches. Figure 4 indicates that there is no apparent difference in the manner water VAPOR transport is managed across these two different underlayment systems. Follow-up simulations showed, if time-aged roof shingles deteriorated and water penetration occurred onto the underlayment, substantially higher moisture accumulation in the OSB occurred for the VAPOR permeable underlayment. Validation Initial testing used 7 16" OSB and Owens Corning Classic 3-tab shingles on each sample, with 3 different underlayments: standard 15-lb. felt, Fiberglas Reinforced Felt and a nonbreathable brand of synthetic underlayment, using three full-system boards for each testing validated the initial computer model, but additionally it drove further testing on the individual components. Part of this was the testing of the shingle overlap layer which was discussed earlier, but additionally it validated the large-scale test specimen developed to facilitate the second question examined was the impact of sample conditioning on the overall test results.